The work presented in this thesis was carried out in the framework of two European projects: SOLWATER, and AQUACAT. The global objective of those projects, including the present work, was to set an autonomous system for drinking water production in developing countries using solar energy. More specifically, the technology of drinking water decontamination by photocatalysis using inexpensive TiO2 catalyst and free of charge sun energy was investigated. In details, the research project dealt with the correlation of the TiO2 physicochemical characteristics with its photocatalytic activity on both organic and bacterial removal, the comparison of the photoactivity of suspended and fixed TiO2, and the development of a large scale prototype for fixed catalyst system. Finally, this work also included the creation of a mathematical model to predict the photo and photocatalytic bacterial inactivation. In photocatalytic experiments, phenolic compounds (phenol, 4-nitrophenol, 4-hydroxybenzoic acid and gallic acid) and benzoic acid were used as model organics, and E. coli as model bacterium. The titania physicochemical characteristics, crystalline structure, specific surface area (BET), particle size, aggregate size, and isoelectric point (IEP) of 13 different suspended TiO2 samples were determined and correlated with photocatalytic activity for chemicals and bacteria abatement. Organics degradation kinetics were found to be significantly affected by the TiO2 surface charge, i.e. the IEP. Indeed, the organic compounds degradation was greatly enhanced by negatively charged TiO2 samples. BET affected differently the primary compounds degradation and the total organic carbon mineralization. The primary degradation kinetics of strongly adsorbed pollutants was enhanced by large BET. Whereas, there was an optimum BET value for a total organic carbon removal, for all organic compounds studied. In addition, the mixed crystalline structure anatase-rutile, in the form of the commercially available Degussa P25, was more active for the purpose of organic degradation than pure anatase TiO2. Bacteria abatement was not influenced by BET specific surface area, particle and aggregate size of suspended TiO2. However, surface charge was crucial for the TiO2 and bacteria interaction. Indeed, negatively charged TiO2 were not efficient in inactivating E. coli, probably due to electrostatic repulsions with the negatively charged E. coli outer membrane. Depending on the nature of the contaminants, different catalyst characteristics have to be considered with regards to TiO2 efficiency on water detoxification and disinfection. Thus, it is not possible to draw a general scheme to determine which photocatalyst is the most suitable for the decontamination of particular water. Comparison of suspended and fixed catalyst revealed that the system with suspended TiO2 is more efficient for bacteria inactivation, yet the post-treatment catalyst filtration presents a serious disadvantage compared to fixed TiO2. Large scale experiments in compound parabolic collector (CPC) prototype reactor with fixed TiO2 revealed that a significant photocatalytic effect appeared only at a low bacterial initial concentration (≤ 103 CFU/ml). High organic compounds removal efficiency with fixed TiO2 was observed at laboratory and large scales, with kinetics similar to the ones obtained with suspended TiO2. A mathematical model was developed to predict bacterial photo-inactivation as a function of the light intensity and of the bacteria physiological state. The modelled curves fitted well with the experimental results for both exponential and stationary phases. Moreover, bacteria were more sensitive to light irradiation in the exponential than in the stationary state. Solar photocatalysis with fixed TiO2 was shown to be an efficient technology for organic pollutant removal and could be used as complement to other treatment process. Indeed, due to the slow bacteria inactivation kinetics observed, the technology needs further development to be used as a unique drinking-water treatment.